Article, Pulmonology

ED visit volume and quality of care in acute exacerbations of chronic obstructive pulmonary disease

Original Contribution

ED visit volume and quality of care in acute exacerbations of chronic obstructive pulmonary disease^?

Chu-Lin Tsai MD, MPH ^a,b,?, Brian H. Rowe MD, MSc ^c,

Rita K. Cydulka MD, MS ^d, Carlos A. Camargo Jr MD, DrPH ^a,b

^aDepartment of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA

^bDepartment of Epidemiology, Harvard School of Public Health, Boston, MA 02115, USA

^cDepartment of Emergency Medicine, University of Alberta, Edmonton, AB, Canada T6G 2B7

^dDepartment of Emergency Medicine, MetroHealth Medical Center, Case Western Reserve University, Cleveland, OH 44109, USA

Received 5 June 2008; accepted 29 July 2008

Abstract

Objective: The purpose of this study is to determine whether emergency department (ED) visit volume is associated with ED quality of care in patients with acute exacerbations of chronic obstructive pulmonary disease (COPD).

Methods: We performed a prospective Multicenter cohort study involving 29 EDs in the United States and Canada. Using a standard protocol, we interviewed consecutive ED patients with COPD exacerbation, reviewed their charts, and completed a 2-week telephone follow-up. The associations between ED visit volume and quality of care (process and outcome measures) were examined at both the ED and patient levels.

Results: After adjustment for patient mix in the multivariable analyses, chest radiography was less frequent among patients with COPD exacerbations in the low-volume (odds ratio [OR], 0.2; 95% confidence interval [CI], 0.1-0.4) and high-volume EDs (OR, 0.1; 95% CI, 0.05-0.5), with medium- volume EDs as the reference. arterial blood gas testing was less frequent in the low-volume EDs (OR, 0.1; 95% CI, 0.02-0.8). Medication use was similar across volume tertiles. With respect to outcome measures, patients in high-volume EDs were more likely to be discharged (OR, 4.2; 95% CI, 2.2-7.7) and to report ongoing exacerbation at a 2-week follow-up (OR, 1.9; 95% CI, 1.02-3.5).

Conclusions: Traditional positive volume-quality relationships did not apply to emergency care of COPD exacerbation. High-volume EDs used less guideline-recommended Diagnostic procedures, had a higher admission threshold, and had a worse short-term patient-centered outcome.

Introduction

^? The present cohort studies were supported by an unrestricted grant from Boehringer Ingelheim (Ridgefield, Conn, and Burlington, Ontario, Canada).

* Corresponding author. EMNet Coordinating Center, Department of

Emergency Medicine, Massachusetts General Hospital, Boston, MA 02114. Tel.: +1 617 726 5276; fax: +1 617 724 4050.

E-mail address: [email protected] (C.-L. Tsai).

A number of studies have demonstrated a positive association between volume of hospital services and patient outcomes, across a wide range of clinical conditions [1]. In other words, institutions with higher visit volumes experi- ence better clinical outcomes compared with institutions with lower visit volume. However, most of the volume-quality

research has focused on surgical outcomes, particularly in the inpatient setting. Little is known about the relationship between volume and quality of care for medical conditions in the emergency setting. Emergency department (ED) visit volumes have risen sharply in recent years, from 93 million visits in 1994 to 115 million visits in 2005 [2]. Given that rising ED visit volumes may strain ED resources and compromise ED quality of care [3,4], it remains unclear if high ED visit volume can represent the concept of “practice makes perfect,” an idea that is believed to underlie traditional volume-quality relationships.

Chronic obstructive pulmonary disease (COPD) exacer- bation is a commonly seen medical disorder in the ED, accounting for approximately 1.5 million ED visits in the United States per year [5]. chronic obstructive pulmonary disease exacerbation also is the third most common cause of hospitalization, with an estimated 726 000 hospitalizations in 2000 [5]. Previous studies have demonstrated important differences between Guideline recommendations and actual management of COPD exacerbation, either in the ED [6] or during hospitalization [7]. Understanding the structural factors (eg, visit volume) that explain the variation in ED practice may guide quality improvement efforts in emer- gency care of COPD exacerbation.

Because of the complex meaning of volume in the context of emergency medicine, we hypothesized that traditional positive volume-quality relationship would not hold in the emergency care of COPD exacerbation. The goal of our analysis was to determine whether ED visit volume is associated with ED quality of care in terms of process and outcome measures in a multicenter study of COPD exacerbation.

Methods

Study population

This observational study combined data from 2 identical prospective cohort studies performed in 2000 to 2001 as part of the Multicenter Airway Research Collaboration, a division of the Emergency Medicine Network (EMNet, www.emnet- usa.org). Details of the study design and data collection have been published previously [6]. In brief, using a standard protocol, investigators at 29 EDs in 15 US states and 3 Canadian provinces provided 24-hour/d coverage for a process measures“>median of 2 weeks. repeat visits by individual subjects were excluded. All patients were managed at the discretion of the treating physician. Inclusion criteria were physician diag- nosis of COPD; presenting to the ED for treatment of COPD exacerbation, as defined by increasing shortness of breath, worsening cough, or change in sputum production [8,9]; 55 years or older; and the ability to give informed consent. The institutional review boards at each of the 29 participating hospitals approved the study, and informed consent was obtained from all participants.

Data collection and processing

Trained research personnel performed the ED interview and assessed patients’ demographic and baseline clinical characteristics, COPD history, and details of their current COPD exacerbation. Data on ED management and disposi- tion were obtained from chart review. follow-up data were collected by telephone interview 2 weeks later. All forms were reviewed by site investigators before submission to the EMNet Coordinating Center in Boston, Mass, where they underwent further review by trained personnel and then double data entry.

Information on patient-level covariates

The preferred term used for a patient’s diagnosis (ie, COPD, emphysema, or chronic bronchitis) was determined by the following question: “Has a doctor ever said that you have asthma, COPD, emphysema, or chronic bronchitis?” Patients were assigned to the COPD group if they reported COPD, emphysema, or chronic bronchitis. Insurance status was categorized as private (commercial or private), Medi- caid, other public (all Canadian and Medicare), or none. Primary care provider status was assigned on the basis of the following question: “Do you have a primary care provider (such as a family doctor, internist, or nurse practitioner)?” Quality of life was measured by the Short-Form Chronic Respiratory Disease Questionnaire (SF-CRQ) [10].

Information on ED-level covariates

Annual ED visit volumes were taken from the National Emergency Department Inventory [11]. National Emergency Department Inventory was developed by staff at the EMNet Coordinating Center and contains data on 4862 US EDs for the year 2001, including annual ED visit volume, facility location, emergency medicine residency program, and urban vs rural status. Geographic regions (northeast, south, midwest, and west) were defined according to Census Bureau boundaries.

Information on whether the ED has a designated COPD treatment area and guidelines for managing COPD was obtained from investigator contact.

Quality indicators

- 1. Process measures

On the basis of recommendations contained in the contemporaneous COPD guidelines [8,9,12,13], we categor- ized the following diagnostic evaluations and treatments as beneficial: chest radiography, arterial blood gas analysis, inhaled short-acting ?-agonists, inhaled anticholinergic bronchodilators, Systemic corticosteroids, antibiotics, and discharge medications with Systemic corticosteroids [14]. The use of bronchodilator and systemic corticosteroids was

recently included as quality measures in pharmacotherapy for acute exacerbation of COPD (AECOPD) in the Health Care Effectiveness Data and Information Sets [15]. To better define the patient population eligible for the process quality indicators, we restricted the denominator of arterial blood gas analysis to admitted patients as appropriate candidates (ie, severe exacerbation) [16]. We also restricted the denominator of antibiotic use to patients with changes in sputum volume or purulence, according to guideline recommendations [17]. Finally, we considered treatment with methylxanthine bronchodilators as nonbeneficial [18].

- 1. Outcomes measures

Emergency department outcomes included length of stay and disposition. Hospital admission was defined as admis- sion to an Inpatient unit, observation unit, or intensive care unit. Follow-up outcomes were relapse and/or ongoing exacerbation. Relapse was defined as a worsening of respiratory symptoms that led to an urgent unscheduled clinic visit or ED visit for further care, including early relapse (within 48 hours) and late relapse (within 2 weeks). Ongoing exacerbation during the 2-week follow-up period was determined by change score in the SF-CRQ [10]. Ongoing exacerbation was assigned to patients whose mean improve- ment in quality of life score did not reach the minimal clinically important difference or who stated that their COPD was “about the same” or worse than at the time of their ED presentation [10]. Because discharge decisions often rely on nuanced physician judgments and do not solely lend to the easy interpretability of quality, we used the follow-up outcomes to help determine the appropriateness of the discharge decision. For example, a higher early relapse rate may indicate the inappropriateness of the discharge decision at the index ED visit.

Statistical analysis

Summary statistics at both the patient and ED level are presented as proportions (with 95% confidence intervals [CIs]) or medians (with interquartile ranges [IQRs]). The 29 participating EDs were grouped into tertiles of ED visit volume: low- (n = 10), medium-(n = 10), and high-volume EDs (n = 9). For categorical variables, the associations between ED visit volume and variables were examined using ?2 or Fisher exact tests, as appropriate. For continuous variables, the associations between ED visit volume and variables were examined using Kruskal-Wallis tests after assessing the data for normality. To adjust for baseline patient characteristics (patient mix) that may have confounded the relationship between ED visit volume and quality indicators, we performed multivariable logistic regression modeling. Age and sex were included throughout the model building process because of their potential clinical significance. Variables associated with ED volume or outcome at P b

.10 in univariate analyses were considered eligible for inclusion in the multivariable model. Variables were retained

in the final model if their inclusion changed the odds ratios (ORs) for the volume tertiles by 10% or more [19]. To account for the effects of clustering of patients within EDs, the multivariable models were fit using generalized estimat- ing equations [20]. The discrimination and calibration of these models was determined by c-statistic and Hosmer- Lemeshow test, respectively [21]. The relationships between annual ED visit volume and risk-adjusted probabilities of quality indicators were plotted, and locally weighted scatterplot smoothed curves were fitted to characterize the dose-response relationship [22].

To assess the potential impact of the Canadian vs US health care system on the results, we repeated analyses after exclusion of all Canadian patients. The results were similar and are not reported. All ORs are presented with 95% CIs. All analyses were performed using Stata v9.0 software (StataCorp, College Station, Tex). All P values are 2 sided, with P b .05 considered statistically significant.

Results

Of the 853 potential COPD and/or asthma patients screened, 584 had physician-diagnosed COPD (mixED asthma-COPD or COPD only). Of these 584 COPD patients, 397 (68%) were enrolled. Enrolled and nonenrolled patients were similar across several sociodemographic factors (data not shown) except age; nonenrolled patients were older than enrolled patients (median age, 73 vs 69 years, P b .001). A 2-week follow-up was obtained for 349 (88%) of the enrolled patients.

Emergency department characteristics

Twenty-four US and 5 Canadian sites participated in this study (Table 1). Overall, the median annual ED visit volume was 60 832 (IQR, 52 250-73 536). The range of the annual number of ED visits was 18 345 to 52 250 for the low- volume EDs, 54 139 to 71 250 for the medium-volume EDs, and 73 536 to 187 798 for the high-volume EDs. Across these volume tertiles, there were no significant differences in the following: hospital ownership, emergency residency program, designated COPD treatment area, or guideline for managing COPD. Among the US sites, there were no differences in geographic location by volume groups.

Patient characteristics

Sociodemographic and baseline clinical characteristics are shown in Table 2. The median age of this cohort was 69 years, and 52% were women. The high-volume EDs were more likely to receive patients who were male, were nonwhite, had lower levels of education, and were uninsured or insured by Medicaid. Patients in the high-volume EDs were more likely to use the ED as their usual site for COPD care, including COPD prescriptions, and had a high number of ED visits in

Median annual volume	43 306 (35 344-47 844)	64 396 (59 495-70 581)	85 069 (75 000-98 253)
Canadian site	1 (10)	2 (20)	2 (22)	.85
Public hospital	2 (20)	3 (30)	6 (67)	.13
Emergency medicine residency program	9 (90)	10 (100)	8 (89)	.75
Has designated COPD treatment area	1 (10)	2 (20)	4 (50)	.21
Has guideline for managing COPD	3 (30)	3 (30)	1 (13)	.75
Region ^a				.23
Northeast	5 (56)	4 (50)	2 (29)
Midwest	2 (22)	2 (25)	1 (14)
South	0 (0)	1 (13)	4 (57)
West	2 (22)	1 (13)	0 (0)
Data are presented as number (percentage) or median (IQR) unless otherwise specified. ^a Among 24 US EDs (low volume, n = 9; medium volume, n = 8; high volume, n = 7).

the past year. There were no significant differences in markers of chronic COPD severity across volume tertiles, except that more patients had been intubated in the high-volume ED group. Comorbidity status did not differ across volume tertiles. Overall, the symptoms of the index COPD exacer- bation and ED presentations (eg, respiratory rate or oxygen saturation) did not differ across volume tertiles. The presence of airway infection symptoms (eg, altered sputum purulence or volume) was also comparable across 3 groups; however, the patients in the high-volume EDs were less likely to be diagnosed as having pneumonia at the index ED visit.

Table 1 Characteristics of 29 participating EDs, according to ED visit volume

Characteristic

ED volume

Low (n = 10)

Medium (n = 10)

High (n = 9)

Quality of care

- 1. Process measures

With respect to ED diagnostic testing, the performance rates of chest radiography and arterial blood gas were lower in the low- and high-volume ED compared with the medium- volume EDs (Table 3). The Prescription rates of Inhaled bronchodilators, systemic corticosteroids, and antibiotics were similar across tertiles. At discharge, patients were equally likely to receive systemic corticosteroids across volume groups. Only one patient in the high-volume group received treatment with methylxanthines in the ED.

- 1. Outcomes measures

More than two thirds of the patients were admitted in the medium-volume EDs, and the length of stay was also higher in this group, with 50% of patients staying in the ED for 6 hours or longer. By contrast, more than half of the patients were discharged from the high-volume EDs. Patients in the high-volume EDs were more likely to have relapse in 48 hours and to report ongoing exacerbation during a 2-week follow-up.

- 1. Multivariable analyses

Multivariable analyses showed that chest radiography was less frequently performed in the low- and high-volume

EDs compared with the medium-volume EDs (Table 4). Arterial blood gas among the admitted was less frequently used in the low-volume EDs. Performance rates for the other process measures pertaining to medication use were similar across volume tertiles (data not shown). With respect to outcome measures, patients in the low-volume EDs had a shorter ED length of stay. Patients in the high-volume EDs were more likely to be discharged and to report ongoing exacerbation at 2-week follow-up. The number of early relapse events (within 48 hours) was too small to perform multivariable modeling. No differences in relapse at 2 weeks were observed across volume tertiles. The final multivariable models showed moderate-to-good discriminative ability, with c-statistics ranging from 0.70 to 0.83. The Hosmer- Lemeshow tests demonstrated a moderate-to-good fit, with P values ranging from .14 to .83. Graphical analyses of dose- response relationship between ED visit volume and quality indicators is shown in Fig. 1. Overall, the use of Processes of care and ED length of stay appeared to increase as visit volume increased but to decrease when volume went beyond a certain level. For example, the performance rates of chest radiography and arterial blood gas increased as visit volume increased but to decrease after a volume cut-point of 60 000 to 70 000. In contrast, the rates of discharge and ongoing exacerbation increased as volume increased.

Discussion

This prospective multicenter study examined the relation- ship between total ED visit volume and quality of care among patients presenting to North American EDs with acute Exacerbations of COPD. Contrary to traditional volume-quality relationships, higher volume did not translate into better quality of care in the emergency setting. The high- volume EDs used less guideline-recommended diagnostic testing [8,9,12,13], and patients in the high-volume EDs were more likely to experience ongoing exacerbation after

Table 2 Demographic and clinical characteristics of ED patients with an acute exacerbation of COPD, according to ED visit volume

Characteristic	Low-volume ED	Medium-volume ED	High-volume ED	P
No. of patients	112	173	112
Sociodemographic factor
Age (y)	69 (62-76)	70 (64-77)	67 (61-74)	.20
Female	58	54	42	.04
Race/ethnicity				b.001
White	70	80	55
African American	24	15	28
Hispanic	4	4	16
Other	3	1	1
high school graduate	53	63	42	.002
Insurance and Health care use Insurance status
Private	18	38	10	b.001
Medicaid	17	9	23
Other public	61	48	43
None	4	5	24
Had primary care provider	92	92	85	.08
ED is usual site for problem COPD care	54	41	68	b.001
ED is usual site for COPD prescriptions	13	10	24	.005
No. of admission for COPD in the past year	0 (0-1)	0 (0-1)	0 (0-2)	.87
No. of ED visits in the past year	1 (0-3)	1 (0-2)	3 (1-6)	b.001
Markers of stable COPD severity
Duration of COPD history (y)	8 (3-20)	8 (3-19)	10 (4-19)	.27
Breathing between COPD exacerbations ^a	2 (2-3)	3 (2-4)	3 (2-4)	.09
Ever admitted for COPD	71	63	63	.39
Ever intubated for COPD Comorbidities	10	10	23	.004
Coronary artery disease	25	19	24	.35
Congestive heart failure	21	18	17	.77
Cancer	10	15	6	.06
Depression	14	12	10	.59
Symptoms of index COPD exacerbation
Duration of symptoms b24 h 31		26	35	.28
Severity of COPD exacerbation in past 24 h ^b
Frequency of COPD symptoms	3 (3-4)	3 (3-4)	3 (3-4)	.83
Severity of COPD symptoms	4 (3-4)	4 (3-4)	3 (3-4)	.06
Activity limitations Airway infectious symptoms ^c	4 (3-4)	4 (3-4)	3.5 (3-4)	.18
More sputum production	52	52	59	.78
Change in sputum color	23	44	34	.18
Fever	48	43	50	.79
ED presentation
Respiratory rate (breaths/min)	24 (22-28)	24 (22-28)	24 (22-28)	.28
O₂ saturation (%) Arterial blood gas ^d	94 (90-96)	93 (88-96)	94 (92-97)	.07
PO2, mm Hg	80 (59-98)	62 (54-79)	63 (51-78)	.38
PCO2, mm Hg	47 (38-61)	46 (39-59)	49 (38-60)	.85
Concomitant medical disorder Pneumonia	17	20	8	.02
Congestive heart failure	11	13	12	.83
Data are presented as percentage or median (IQR) unless otherwise specified. ^a On 4-point ordinal scale (1, no symptoms; 2, some symptoms on some days; 3, some symptoms on most days; 4, symptoms most of the time). ^b On a 4-point ordinal scale (1, none; 2, mild; 3, moderate; 4, severe). ^c Available among a subset of patients (n = 122). ^d Available among a subset of patients (n = 113).

Variable	Low-volume ED	Medium-volume ED	High-volume ED	P
No. of patients Process measure Beneficial Chest radiography Arterial blood gas among the admitted ^a Inhaled ?-agonist medication in ED Inhaled anticholinergic medication in ED Received systemic corticoSteroid treatment in ED Received antibiotics in ED ^b Sent home on systemic corticosteroids ^c Nonbeneficial Received methylxanthines in ED Outcome measure ED outcomes ED disposition Hospital admission Discharge from ED Other (eg, left against medical advice) ED length of stay N6 h Post-ED outcomes ^d Relapse within 48 h ^e Relapse within 2 wk ^e Ongoing exacerbation ^f	112	173	112
	86	97	73	b.001
	40	72	56	.002
	94	89	91	.40
	78	73	82	.22
	61	64	60	.86
	50	30	14	.07
	55	56	62	.71
	0	0	1	.57
				b.001
	61	73	39
	37	26	59
	2	1	2
	28	50	35	.001
	0	2	7	.02
	18	13	23	.15
	28	36	50	.008
Data are presented as percentage or median (IQR) unless otherwise specified. ^a Restricted to admitted patients (n = 234). ^b Restricted to patients with altered sputum characteristics (n = 79). ^c Restricted to patients sent home from ED (n = 151). ^d Restricted to patients available for follow-up (n = 349). ^e Relapse event based on patient reporting a worsening of COPD symptoms that led to an urgent care visit. ^f See Methods section for details.

discharge. There also was some evidence that low-volume EDs used less guideline-recommended diagnostic testing [8,9,12,13]. Taken together, the ED quality of care of COPD exacerbations appears to improve as visit volume increases but to decrease when volume goes beyond a certain level. In other words, an inverted U-shaped relationship was demon- strated between visit volume and ED quality of care.

Table 3 Quality of care in ED patients with an acute exacerbation of COPD, according to ED visit volume

The positive linear relationships between processes of care and volume have been challenged by a few studies in other settings. Similar to ours, a recent study found an inverted U-shaped relationship between hospital volume and processes of cardiovascular care [23]. Another study of patients hospitalized for COPD found no association between higher annual volume of admissions for COPD

Table 4 Multivariable risk-adjusted performance on selected quality indicators in ED patients with an acute exacerbation of COPD, according to ED visit volume
	Quality Indicators ^a	Adjusted OR (95% CI) ^b
		Low-volume ED	Medium-volume ED	High-volume ED
	Process measure Chest radiography	0.2 (0.1-0.4)	1.0 (referent)	0.1 (0.05-0.5)
	Arterial blood gas among admitted Outcome measure	0.1 (0.02-0.8)	1.0 (referent)	0.3 (0.1-2.5)
Length of ED stay N6 h		0.3 (0.1-0.8)	1.0 (referent)	0.6 (0.3-1.6)
Discharge from ED		1.6 (0.7-3.4)	1.0 (referent)	4.2 (2.2-7.7)
Post-ED ongoing exacerbation		1.0 (0.5-2.1)	1.0 (referent)	1.9 (1.02-3.5)
^a There were no statistically significant differences in the adjusted ORs of the following quality indicators: inhaled short-acting ?-agonists, inhaled anticholinergic bronchodilators, systemic corticosteroids, antibiotics, and discharge medications with systemic corticosteroids, and relapse within 2 weeks. ^b Values were obtained by logistic regression model with generalized estimating equations and were adjusted for age, sex, race/ethnicity, insurance status, high school graduate, history of intubation for COPD, number of ED visits in the past year, ED as usual site for problem COPD care, ED as usual site for COPD prescriptions, activity limitations in the past 24 hours, respiratory rate at ED presentation, and oxygen saturation.

Fig. 1 Selected risk-adjusted probabilities of quality indicator performance in ED patients with an acute exacerbation of COPD, according to ED visit volume. Quality indicators include 2 process measures (panels A and B) and 3 outcome measures (panel C, D, and E). The fitting lines represent locally weighted scatterplot smoothed curves. CXR indicates chest radiography; ABG, arterial blood gas; LOS, length of stay.

and improved hospital performance of guideline-recom- mended care [7]. Our study extends the volume-quality relationship in acute exacerbations of COPD to a previously unstudied ED population. Moreover, there is evidence suggesting high volume may have negative impact on quality of care. A study of pneumonia demonstrated an inverse association between hospital volume and antibiotic timing [24]. The authors reasoned that the benefit associated with institutional experience may be outweighed by the ED

crowding [24]. Given these recent findings, the traditional view of higher volume as a determinant of better quality of care may need to be more rigorously examined and carefully interpreted in light of both the disease and the specific clinical setting.

Why were the low- and high-volume EDs less likely to use beneficial diagnostic testing, such as chest radiography? Although the underlying mechanisms are not fully clear, low-volume EDs may have less staff and resources for

dealing with COPD exacerbations. A previous study has shown that low-volume EDs have 2-fold higher odds of missed acute myocardial infarctions compared with higher- volume ones [25]. At the other end of volume spectrum, the larger number of patients may have encouraged physicians to adopt a less-aggressive test ordering approach to expedite patient flow. This approach may lead to underdiagnosis of pneumonia and other COPD mimics [26] in the high-volume EDs, as suggested by our data.

Another distinct practice pattern in the high-volume EDs was their relatively short ED length of stay, despite their high visit volume. There are several plausible explanations for this finding. First, less diagnostic evaluation (eg, chest radio- graphy) was performed in the high-volume EDs, resulting in less radiology and laboratory turnaround time [27]. Second, even after adjustment for patient mix, more patients were discharged from the high-volume EDs. It is possible that physicians working in the high-volume EDs may subcon- sciously raise their admission threshold for these relatively dischargeable patients in an attempt to decrease the ED boarding load [27]. Both could help explain the quick patient flow in the presence of high visit volume.

Although the high-volume EDs appeared to be efficient in this regard, their efficiency may compromise other aspects of quality measures, such as patient-centered outcomes [28]. As our data suggested, patients discharged from the high-volume EDs were more likely to experience early relapse or to report ongoing exacerbation, even after adjustment for patient mix across EDs. One possibility is that they were discharged inappropriately because of a higher admission threshold in the high-volume EDs. Alternatively, a significant portion of these patients discharged from the high-volume EDs were uninsured, and they might have difficulties gaining access to follow-up care [29]. Despite standard practice of discharge medica- tions given in the high-volume EDs, the lack of adequate outpatient follow-up may still lead patients to return to the ED for relapse events or for treatment of lingering exacerbation symptoms.

Our study has several strengths. First, unlike most prior studies on volume-outcome relationship that have used administrative data [7,24] our study obtained data by patient interview supplemented by medical record review. This allowed us to capture more precise data on diagnostic testing and medications and to include patient-centered outcome measures. Second, fewer studies have used multivariable techniques or have accounted for clustering when examining the independent effects of volume. Failure to account for the clustering may lead to smaller P values and falsely positive conclusions [30].

This study has some potential limitations. First, we used annual total ED visit volume instead of annual visit volume of acute exacerbation of COPD. This would somewhat attenuate the positive influence of high volume (institutional experience) on ED quality of care for COPD. However, information on total visit volume is easier to obtain and is

more relevant for policymakers to target quality improve- ment efforts. Second, we have used ED crowding to explain some of our findings, but we did not have detailed metrics of ED crowding, such as ED staffing ratios. However, our goal is to characterize the complex volume effect in the emergency setting, not crowding per se, and the findings are directly relevant to our objective. Third, the ED sample is not representative of all North American EDs. It is a volunteer convenience sample of ED sites that overrepre- sents the academic urban setting. It is possible that there could be more variation in quality of care if we were able to include truly smaller EDs and/or nonacademic EDs. Finally, compared with other ED conditions, such as pneumonia or acute myocardial infarction, the standardized quality indica- tors for acute exacerbations of COPD has just launched [15]. We have attempted to incorporate more guideline-recom- mended process measures and important patient-centered outcome measures. Some of the quality measures we used had limited statistical power because of small number of cases, such as methylxanthine use and early relapse.

Conclusions

In summary, our results suggest that the traditional positive volume-quality relationships did not apply to emergency care for COPD exacerbations. High-volume EDs used less guideline-recommended diagnostic proce- dures, adopted a higher admission threshold, and had a worse short-term patient-centered outcome. Given that the demand for emergency care will likely increase with an aging and growing population, our study has important implications for health services researchers and policymakers. For research- ers, our study represents the first step toward unraveling the complexity of volume-quality relationships in a previously unstudied ED population. Further studies are needed to examine the mechanisms through which total visit volume might influence ED quality of care and whether the observed associations hold for other common ED presentations. For policymakers, the logical next step would be to implement quality improvement initiatives to reduce variation in practice across EDs and removal of the factors that are adversely affecting quality of care.

Acknowledgments

The authors thank the EMNet investigators for their ongoing dedication to public health research, with an emphasis on the treatment and prevention of respiratory/ allergy emergencies. The present cohort studies were supported by an unrestricted grant from Boehringer Ingel- heim (Ridgefield, Conn, and Burlington, Ontario, Canada).

EMNet Steering Committee:

Edwin D. Boudreaux, PhD; Carlos A. Camargo Jr, MD (Chair); Jonathan M. Mansbach, MD; Steven Polevoi, MD;

Michael S. Radeos, MD, MPH; and Ashley F. Sullivan, MPH, MS.

EMNet Coordinating Center:

Angela T. Anderson; Carlos A. Camargo Jr, MD (Director); Lisa A. Dubois; Janice M. Espinola, MPH; Jessica M. Pang; Ashley F. Sullivan, MS, MPH; and Chu-Lin Tsai, MD, MPH-all at Massachusetts General Hospital, Boston, Mass.

Principal Investigators at the 29 Participating Sites:

FC Baker III (Maine Medical Center, Portland, Me); MP Blanda (Summa Health System, Akron, Ohio); ED Bou- dreaux (Earl K. Long Memorial Hospital, Baton Rouge, La); BE Brenner (The Brooklyn Hospital Center, Brooklyn, NY); CA Camargo Jr (Massachusetts General Hospital, Boston, Mass); RK Cydulka (MetroHealth Medical Center, Cleve- land, Ohio); TJ Gaeta (New York Methodist Hospital, Brooklyn, NY); B Goldfeder (Shands Hospital at the University of Florida, Gainesville, Fla); RJ Grant (Hartford Hospital, Hartford, Conn); RO Gray (Hennepin County Medical Center, Minneapolis, Minn); A Guttman (Sir Mortimer B. Davis-Jewish General Hospital, Montreal, QC); LW Kreplick (Christ Hospital and Medical Center, Oak Lawn, Ill); DS Mackey (Lethbridge Regional Hospital, Lethbridge, AB); A Mangione (Albert Einstein Medical Center, Philadelphia, Pa); J Peters (University of Texas Health Sciences Center at San Antonio, San Antonio, Tex); MS Radeos (Lincoln Medical Center, Bronx, NY); PL Rice (Brigham and Women’s Hospital, Boston, Mass); BH Rowe (University of Alberta Hospital, Edmonton, AB); M Sama (St Joseph Mercy Hospital, Ann Arbor, Mich); D Schreiber (Stanford University Medical Center, Stanford, Calif); NI Shapiro (Beth Israel Deaconess Medical Center, Boston, Mass); PC Shukla (University of Texas Southwestern Medical Center, Dallas, Tex); D Sinclair (Queen Elizabeth II Health Science Centre, Halifax, NS); H Smithline (Baystate Medical Center, Springfield, Mass); PE Sokolove (UC Davis Medical Center, Sacramento, Calif); M Steffens (Palmetto Richland Memorial Hospital, Columbia, SC); CA Terregino (Cooper Hospital/University Medical Center, Camden, NJ); A Travers (Royal Alexandria Hospital, Edmonton, AB); and EJ Weber (UCSF Medical Center, San Francisco, Calif).

References

Halm EA, Lee C, Chassin MR. Is volume related to outcome in health care? A systematic review and methodologic critique of the literature. Ann Intern Med 2002;137(6):511-20.
National Center for Health Statistics (U.S.). National Hospital Ambulatory Medical Care Survey. Emergency department summary. Hyattsville (Md): U.S. Dept. of Health and Human Services Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics, DHHS publication; 2004.
Derlet RW, Richards JR. Overcrowding in the nation’s emergency departments: complex causes and disturbing effects. Ann Emerg Med 2000;35(1):63-8.
Institute of Medicine (U.S.). Committee on the Future of Emergency Care in the US Health System.: the future of emergency care in the United States health system. Washington (DC): National Academy Press; 2006.
Mannino DM, Homa DM, Akinbami LJ, Ford ES, Redd SC. Chronic obstructive pulmonary disease surveillance-United States, 1971- 2000. MMWR Surveill Summ 2002;51(6):1-16.
Cydulka RK, Rowe BH, Clark S, Emerman CL, Camargo Jr CA. Emergency department management of acute exacerbations of chronic obstructive pulmonary disease in the elderly: the Multicenter Airway Research Collaboration. J Am Geriatr Soc 2003;51(7):908-16.
Lindenauer PK, Pekow P, Gao S, Crawford AS, Gutierrez B, Benjamin EM. Quality of care for patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease. Ann Intern Med 2006;144 (12):894-903.
Celli BR, MacNee W. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 2004;23(6):932-46.
Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic obstructive lung disease (GOLD) Workshop summary. Am J Respir Crit Care Med 2001;163(5):1256-76.
Tsai CL, Hodder RV, Page JH, Cydulka RK, Rowe BH, Camargo Jr CA. The Short-Form Chronic Respiratory Disease Questionnaire was a valid, reliable, and responsive quality-of-life instrument in acute exacerbations of chronic obstructive pulmonary disease. J Clin Epidemiol 2008;61(5):489-97.
Sullivan AF, Richman IB, Ahn CJ, et al. A profile of US emergency departments in 2001. Ann Emerg Med 2006;48(6):694-701.
Snow V, Lascher S, Mottur-Pilson C. Evidence base for management of acute exacerbations of chronic obstructive pulmonary disease. Ann Intern Med 2001;134(7):595-9.
Bach PB, Brown C, Gelfand SE, McCrory DC. Management of acute exacerbations of chronic obstructive pulmonary disease: a summary and appraisal of published evidence. Ann Intern Med 2001;134(7):600-20.
Aaron SD, Vandemheen KL, Hebert P, et al. Outpatient oral prednisone after emergency treatment of chronic obstructive pulmonary disease. N Engl J Med 2003;348(26):2618-25.
National Committee on Quality Assurance. The Health Care Effectiveness Data and Information Sets (HEDIS), 2008 quality measures for pharmacotherapy of COPD Exacerbation. Available at: http://www.ncqa.org/Portals/0/HEDISQM/HEDIS2008/2008_ Measures.pdf [Accessed June 5, 2008].
Rodriguez-Rosin R. Toward a consensus definition for COPD exacerbations. Chest 2000;117(5 suppl 2):398S-401S.
Rabe KF, Hurd S, Anzueto A, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 2007;176(6):532-55.
Barr RG, Rowe BH, Camargo CA. Methylxanthines for exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2003(2):CD002168.
Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989;79(3):340-9.
Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42(1):121-30.
Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: Wiley; 2000 [Wiley series in probability and statistics].
Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology 1995;6(4):356-65.
Williams SC, Koss RG, Morton DJ, Schmaltz SP, Loeb JM. case volume and hospital compliance with evidence-based processes of care. Int J Qual Health Care 2008;20(2):79-87.
Lindenauer PK, Behal R, Murray CK, Nsa W, Houck PM, Bratzler DW. Volume, quality of care, and outcome in pneumonia. Ann Intern Med 2006;144(4):262-9.
Schull MJ, Vermeulen MJ, Stukel TA. The risk of missed diagnosis of acute myocardial infarction associated with emergency department volume. Ann Emerg Med 2006;48(6):647-55.
Emerman CL, Cydulka RK. Evaluation of High-yield criteria for chest radiography in acute exacerbation of chronic obstructive pulmonary disease. Ann Emerg Med 1993;22(4):680-4.
Asplin BR, Magid DJ, Rhodes KV, Solberg LI, Lurie N, Camargo Jr CA. A conceptual model of emergency department crowding. Ann Emerg Med 2003;42(2):173-80.
Institute of Medicine (U.S.). Committee on Quality of Health Care in America. Crossing the quality chasm: a new health system for the 21st century. Washington (DC): National Academy Press; 2001.
Tsai CL, Griswold SK, Clark S, Camargo Jr CA. Factors associated with frequency of emergency department visits for chronic obstructive pulmonary disease exacerbation. J Gen Intern Med 2007.
Panageas KS, Schrag D, Riedel E, Bach PB, Begg CB. The effect of clustering of outcomes on the association of procedure volume and surgical outcomes. Ann Intern Med 2003;139(8):658-65.

AJEM